Question

我试图绘制一条线图，该线图在x中具有两个x-axis变量，在y中具有一个连续的y-axis变量。 x1和x2的计数不同。 df如下所示-

df <- structure(list(val = c(3817,2428,6160,6729,7151,7451,6272,7146,7063,6344,5465,6169,7315,6888,7167,6759,4903,6461,7010,7018,6920,3644,6541,31862,31186,28090,28488,29349,28284,25815,23529,20097,19945,22118), type = c("1wt", "1wt", "3wt", "3wt", "3wt", "5wt", "5wt", "7wt", "7wt", "7wt","10wt","10wt","10wt","15wt","15wt","20wt","20wt","25wt","25wt","25wt","30wt","30wt","30wt","20m","20m","15m","15m","15m","10m","10m","5m", "5m", "5m", "5m"), group = c("A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B")), row.names = c(NA, 34L), class = "data.frame")

其中x变量是-

x1 <- factor(df$type, levels = c("1wt", "3wt", "5wt", "7wt", "10wt", "15wt", "20wt", "25wt", "30wt")) 和

x2 <- factor(df$type, levels = c("20m", "15m","10m","5m"))

我想分别为x1和x2用不同的颜色和图例的行，如x轴上的df$group和y处的df$val-轴。你能帮我做这个吗？预先感谢。

Answer 1

编辑：添加在下面

这是一种假设方法是将A组中可能的类型值的范围与B组中可能的值的范围进行映射。

可以手动添加标签，但是我认为没有一种简单的方法可以在一个图中同时使用两个分类x轴。

df2 <- df %>%
  mutate(x = case_when(type == "1wt" ~ 0,
                       type == "3wt" ~ 1,
                       type == "5wt" ~ 2,
                       type == "7wt" ~ 3,
                       type == "10wt" ~ 4,
                       type == "15wt" ~ 5,
                       type == "20wt" ~ 6,
                       type == "25wt" ~ 7,
                       type == "30wt" ~ 8,

                       type == "20m"  ~ 0/3 * 8,
                       type == "15m"  ~ 1/3 * 8,
                       type == "10m"  ~ 2/3 * 8,
                       type == "5m"   ~ 3/3 * 8))


ggplot(df2, aes(x, val, color = group, group = group)) + 
  geom_point() +
  geom_smooth(method = lm)

第二种方法

听起来OP希望以某种方式在数字上使用type值。如果它们不是按照所描述的方式在本质上相互链接，那么我怀疑将它们像绘制一样会产生误导。（请参阅here，以了解为什么会造成麻烦。）

也就是说，这就是您的操作方法。首先，这是一种仅使用type的数字部分的方法。请注意，与B组相关联的“ m”在底部，而“ wt”在顶部与组A关联，如以下OP注释中添加的示例所示。我已经在轴上添加了颜色以阐明这一点。视觉上有点违反直觉，因为与上轴相关的点在底部，反之亦然。

df2 <- df %>%
  # First, let's take the number used in "type" without adjustment
  mutate(x_unadj = parse_number(type))

ggplot(df2, aes(x_unadj, val, color = group, group = group)) + 
  geom_point() +
  geom_smooth(method = lm) + # Feel free to use other smoothing method, but
                             # not obvious to me what would be improvement.
  scale_x_continuous("m", sec.axis = sec_axis(~., name = "wt")) +
  theme(axis.text.x.bottom  = element_text(color = "#00BFC4"),
        axis.title.x.bottom = element_text(color = "#00BFC4"),
        axis.text.x.top     = element_text(color = "#F8766D"),
        axis.title.x.top    = element_text(color = "#F8766D"))

如果不能令人满意，我们可以使用

反转两个轴的顺序

scale_x_reverse("m", sec.axis = sec_axis(~., name = "wt")) +

使用ggplot 3.1.0（自2018年10月起），我无法使辅助x轴沿与主轴相反的方向移动。 2017年的This example似乎不再适用于此版本。截至2018年12月，proposed fix正在接受审核，旨在解决此问题。

在单个绘图中为一个连续的Y变量绘制多个X变量

1 个答案: